The leaktightness of a number of vessels was determined at subatmospheric pressure. For structural reasons, the maximum allowable pressure difference between the contained volume and the atmosphere was limited to 0.2 kg/cm2. Inleakage was determined by measuring the oxygen contained in the inleaking air. The oxygen initially present in the vessels was removed in advance by catalytic combination with hydrogen. With this method, the change in the oxygen partial pressure is more perceptible than a change in the absolute pressure. Additionally, a homogeneous temperature distribution inside the vessel is not required. The oxygen partial pressure was measured by an electrochemical cell. A leakage rate of 1.1 ± 0.2 liters/h into a contained volume of 152 m3 (i.e., <0.02%/day) was determined within 41 h after the oxygen was removed. Although the method is accurate and fast and can be used for a reactor containment, even with the reactor in operation, it is applicable only to leakage rate measurements at subatmospheric pressures. It must be pointed out that a leak at subatmospheric pressure may behave quite differently when the internal pressure is greater than the external.